On Sun, Sep 30, 2018 at 7:36 AM Eric Sunshine <[email protected]> wrote:
>
> On Sat, Sep 29, 2018 at 3:11 PM Nguyễn Thái Ngọc Duy <[email protected]>
> wrote:
> > Signed-off-by: Nguyễn Thái Ngọc Duy <[email protected]>
> > ---
> > diff --git a/Documentation/git-reflog.txt b/Documentation/git-reflog.txt
> > @@ -72,6 +72,11 @@ Options for `expire`
> > +--single-worktree::
> > + By default when `--all` is specified, reflogs from all working
> > + trees are processed. This option limits the processing to reflogs
> > + from the current working tree only.
>
> Bikeshedding: I wonder if this should be named "--this-worktree" or
> "--this-worktree-only" or if it should somehow be orthogonal to --all
> rather than modifying it. (Genuine questions. I don't have the
> answers.)
It follows a precedent (made by me :p) which is rev-list
--single-worktree. I doubt that option is widely used though so we
could still rename it if there's a better name. I made
--single-worktree to contrast "all worktrees" by default. Even if it's
"this/current worktree" it still has to somehow say "everything in
this worktree" so I felt modifying --all was a good idea.
> > diff --git a/builtin/reflog.c b/builtin/reflog.c
> > @@ -577,10 +585,18 @@ static int cmd_reflog_expire(int argc, const char
> > **argv, const char *prefix)
> > if (do_all) {
> > struct collect_reflog_cb collected;
> > + struct worktree **worktrees, **p;
> > int i;
> >
> > memset(&collected, 0, sizeof(collected));
> > - for_each_reflog(collect_reflog, &collected);
> > + worktrees = get_worktrees(0);
> > + for (p = worktrees; *p; p++) {
> > + if (!all_worktrees && !(*p)->is_current)
> > + continue;
> > + collected.wt = *p;
> > + for_each_reflog(collect_reflog, &collected);
> > + }
> > + free_worktrees(worktrees);
>
> Should this have a test in the test suite?
Of course. I was partly lazy/tired near the end, and anticipated more
comments anyway so I did not do it :D
--
Duy
